This revised competing renewal application is for continued follow-up and data analyses on a large epidemiologic cohort, which was established in Hawaii and Los Angeles during the period 1993-1996. The population- based cohort, comprised of more than 215,000 members is unique in its multiethnic composition, including substantial numbers of Latinos, African Americans, Japanese Americans, and whites. At entry, each participant complete a 26-page mail questionnaire that contained a quantitative diet history, medical, medication, physical activity, and female reproductive histories; and demographic information. In addition to maintaining a high rate of follow-up on the cohort, we will study the relationship of several dietary factors to four common cancer sites: prostate, breast, colorectum and lung. Associations of these cancers with nutrients (e.g., prostate cancer with saturated fat, lycopene; breast cancer with the ratio of monounsaturated to saturated/ polyunsaturated fat, components of dietary fiber; colorectal cancer with fat, energy, folate; lung cancer with specific fats, carotenoids) and with foods (e.g., prostate cancer with red meat, legumes; breast cancer with high-fiber vegetables; colorectal cancer with meats cooked at high temperature legumes; lung cancer with animal products, food sources of carotenoids) will be examined, taking advantage of both the diversity and range of intakes among cohort members. These relationships will first be examined within each ethnic group. Then, the consistency of relationships among the different ethnic group will be evaluated, using calibrated dietary exposure values based on 24-hour dietary recall data collected on a large subsample of the cohort. Finally, the extent to which dietary and non-dietary data can account for interethnic differences in cancer risk will be assessed. Passive follow-up on the cohort will use computer linkage to the population-based cancer registries in Hawaii and California, and should yield at least 2542 breast, 3426 prostate, 2737 colorectal, and 2034 lung cancer incident cases by the year 2002. Active follow-up will include the administration of a brief follow-up questionnaire in the first two years of the renewal period, as well as regular mailings of a study newsletter. Findings from this study should help to elucidate relationships of diet to cancer, and to better understand the basis for ethnic variations in cancer incidence.